Preview unavailable

🎯 Start here

• Self-assessment exercise - questions.pdf

• Self-assessment exercise - solutions.pdf

👋🏼 Introductions, History of ML, and Course Overview

• lec 01 - handout.pdf

• Welcome 🤗

• Who are we?

• What is intelligence?

• What is machine learning?

• What machine learning is not

• Types of machine learning

• A brief history of machine learning

• An overview of machine learning applications

• Timetable and course logistics

• Quiz

🏘️ Nearest Neighbors

• lec 02 - handout.pdf

• Overview

• What is data?

• Redundancies in data

• How is data represented?

• Datasets of training and testing

• Overview of the nearest neighbors algorithm

• L-p norms and cosine similarity

• The nearest neighbor algorithm

• The k-nearest neighbors algorithm

• Cross-validation & generalization

• On the pitfalls of k-NN & summary2

• Quiz

🌲 Decision Trees

• lec 03 - handout.pdf

• Introduction

• An overview of decision trees

• On the decision boundaries of decision trees

• Discrete decision trees

• On the expressiveness of decision trees

• Intuition for learning decision trees

• Uncertainty quantification, entropy, & information gain2

• Learning decision trees in action

• Summary of decision trees

• Quiz

🌐 Generalization & Ensembling

• lec 04 - handout.pdf

• Motivation

• The bias-variance tradeoff

• Ensembling overview & discussion1

• Bagging

• Boosting

• Summary & discussion

• Quiz

📈 Linear Regression

• lec 05 - handout.pdf

• Motivation

• A modular approach to learning

• Regression step #1 - defining the model

• Loops vs vectorized code

• Regression step #2 - defining the loss

• Regression step #3 - optimizing the loss

• Analytical solution to regression

• Iterative solution to regression (gradient descent)

• Polynomial and regularized regression

• Summary & discussion

• Quiz

🕵🏼‍♂️ Linear Classification & Perceptron

• lec 06 - handout.pdf

• Motivation

• Formulating linear classifiers

• Loss functions for linear classification

• Precision and recall

• The Perceptron algorithm

• On the existence and identification of nonlinear transformations

• Multi-class classification + discussion

• Quiz

〰️ Logistic Regression

• lec 07 - handout.pdf

• Review of linear regression and linear classification

• Linear classification via thresholded activations

• Linear classification via sigmoidal activations

• Summary of attempts to linear classification

• Probabilistic interpretation of logistic regression

• Maximum Likelihood Estimation for logistic regression

• Summary & discussion

• Quiz

📏 Support Vector Machines

• lec 08 - handout.pdf

• Introducing Bernhard Schölkopf

• Motivation (what makes a good linear classifier?)

• Intuition for max-margin classification

• Formulating max-margin classification

• Duality and Lagrangian multipliers

• Learning linear SVMs

• SVMs with slack variables

• Nonlinear SVMs

• Summary

• Quiz

🧠 Neural Networks

• lec 09 & 10 - handout.pdf

• Motivation

• On the linear separability of NOT, AND, & XOR logic gates

• Hand-designed feature engineering

• Biological, computational, and artificial neurons

• Feed-forward neural networks

• Hierarchical feature learning

• Discussion

• Quiz

🔄 Backpropogation

• lec 09 & 10 - handout.pdf

• Mid-class review

• On the expressivity of neural networks

• Solving XOR using neural networks

• An overview of backpropogation

• Backpropogation example #1

• Multivariate chain rule

• Backpropogation example #2

• Putting it all together: the backpropogation algorithm

• Discussion

• Quiz

📷 Convolutional Neural Networks - part I

• lec 11 - handout.pdf

• Introduction

• Neural networks for computer vision

• 1D convolutions

• 2D convolutions

• An intuition for convolutions

• Mid-class review

• Convolutions as layers

• Example convolutions & Toeplitz matrices

• On the hyperparameters of convolutions hparams

• Pooling layers

• 1x1 convolutions and their usecases

• Quiz

📸 Convolutional Neural Networks - part II

• lec 12 - handout.pdf

• Motivation

• Data for computer vision

• Compute for computer vision

• Architectures for computer vision

• Metrics for computer vision

• Training tips & tricks

• Discussion

• Quiz

💭 Natural Language Processing - part I

• lec 13 - handout.pdf

• Motivation

• N-gram language models

• Distributional embeddings

• Neural n-gram models

• Evaluations

• Discussion

• Quiz

💬 Natural Language Processing - part II

• lec 14 - handout.pdf

• Review & overview

• Recurrent Neural Networks example #1

• Recurrent Neural Networks example #2

• Recurrent Neural Networks over words & characters

• Exploding and vanishing gradients

• Long Short-Term Memory models (LSTMs)

• Chaining recurrent cells

• Transformers and their applications

• Summary

• Quiz

🌤️ Probabilistic Models: Discriminative vs Generative

• lec 15 - handout.pdf

• Introduction

• Probability vs likelihood

• Maximum Likelihood Estimation for Bernoullis

• Maximum Likelihood Estimation for Gaussians

• Maximum Likelihood Estimation in past models

• Discriminative vs generative models

• An overview of Bayes classifiers

• On the risk of a classifier

• Gaussian discriminant analysis intuition

• Gaussian discriminant analysis learning, inference, and decision boundaries

• Naive Bayes & summary

• Quiz

🔗 Clustering and Density Estimation: k-Means and Mixtures of Gaussians

• lec 16 - handout.pdf

• Motivation

• K-means overview

• K-means for segmentation & quantization

• K-means objective, loss, and challenges

• Soft K-means

• Mixture of Gaussians intuition

• Mixture of Gaussians in higher dimensions

• Mixture of Gaussians formalities

• Expectation Maximization & discussion

• Quiz

🔥 Dimensionality Reduction: Principal Component Analysis

• lec 17 - handout.pdf

• Motivation

• Formalizing dimensionality reduction

• Vectors, spaces, norms, orthogonality, and independence

• Matrices as linear transformations

• Eigenvalues and eigenvectors

• Eigendecomposition of the covariance matrix

• Step-by-step Principal Component Analysis (PCA)

• Two interpretations of PCA

• Applications of PCA

• Random projections & discussion

• Quiz

🤖 Reinforcement Learning

• lec 18 - handout.pdf

• Motivation for Reinforcement Learning (RL)

• Formulating RL - part 1 (policies)

• Formulating RL - part 2 (value functions)

• Examples of RL systems

• Formulating and solving example MDPs

• Solving MDPs: value iteration and policy iteration

• Exploration & exploitation & discussion

• Quiz

• Student Course Perceptions & class photo

✍️ Self-guided Assignments

• Assignment 1/3.pdf

• Assignment 2/3.pdf

• Assignment 3/3.pdf